2026-09-01

Acquire, Repair, Preserve: A Diagnosis-Guided Post-Training Recipe for Small-Model Dialogue Game Agents

👍 2

08/28 08:00

Interactive dialogue games test a capability that static benchmarks largely leave implicit: a model must carry state across turns, interpret feedback, and choose valid actions under changing constraints. We study this setting in the LM Playschool Challenge with a 2B open-weight model, and find that

EvoUndo: Recoverability-Constrained Self-Evolution for LLM Agent Harnesses

👍 5

08/28 08:00

LLM agents increasingly modify their own prompts, tools, middleware, resources, and execution harnesses at runtime. Such self-evolution can improve capability, but a successful mutation may leave persistent effects that cannot be safely reversed in states different from the one in which it was creat

Sliding-window beats linear attention

👍 11

08/28 08:00

Due to the nature of quadratic attention, Large Language Models (LLMs) consume a lot of memory and energy. Every new token costs more than the previous one. For each additional token, the keys and values must be stored in memory indefinitely, which is unsustainable. Several alternatives have been

LoopArena: Benchmarking Models as Runtime Controllers for Loop Engineering

👍 82

08/28 08:00

Loop Engineering is emerging as a practice for organizing development work around coding agents. Instead of writing each prompt by hand, practitioners design loops that monitor progress, assign work, run checks, and decide what the agent should do next. Even with a capable coding agent, a loop may t

Agentic Artifact Creation: Systems, Evaluation, Principles, and Opportunities

👍 51

08/28 08:00

Generative models can turn natural-language prompts into images, text, code, and other content, lowering the cost of producing drafts and components. Their practical impact increasingly depends on whether those pieces can become complete, dependable deliverables. This survey examines agentic artifac

Locate Anything in Videos: Rethinking Efficient Generative Spatio-Temporal Video Grounding

👍 15

08/28 08:00

Spatio-temporal video grounding (STVG) requires models to identify when a referred event occurs and localize the target entity throughout that interval. Existing multimodal large language models typically serialize dense localization trajectories autoregressively, causing decoding latency to grow wi

Rubric-to-Code Credit Assignment for Reinforcement Learning

👍 5

08/28 08:00

Interactive web application generation requires models to produce usable HTML, CSS, and JavaScript applications from natural language requests. Unlike conventional code generation, application quality depends on multiple user-facing functional requirements, each often tied to localized code regions

Video Generative Models as Geometry Learner

👍 8

08/28 08:00

Recent generative approaches to geometry estimation adapt pretrained image diffusion models and treat the task as image-conditioned generation. Leveraging off-the-shelf image diffusion models, they either (i) train task-specific geometry models (for depth and surface normal estimation) independently

Blind Men and the Elephant: Probing the Epistemic Myopia of LLMs under Long-Tail Divergent Knowledge

👍 15

08/28 08:00

Factual question answering (QA) typically assumes a single canonical answer, obscuring whether large language models (LLMs) retain divergent accounts of long-tail facts. To address this gap, we introduce ElephantBench, a closed-book knowledge probe comprising 1,094 questions generated through an aud

ContextPilot: Teaching Agents for Proactive Context Management via Fine-grained RL

👍 24

08/28 08:00

Long-horizon agentic tasks require large language models (LLMs) to iteratively retrieve, integrate, and maintain dispersed information across multi-turn interactions, but preserving all interaction histories leads to a continuously growing working context. Recent proactive context management methods

Puro-2B: Poor Lab's Qwen2-1.5B Trained on RTX 5090 within $5090

👍 27

08/27 08:00

Language model pretraining has become almost synonymous with prohibitive cost, placing it out of reach for much of the academic and open-source communities. Although strong open-source efforts already exist, including open-weight models and open-source training recipes, a cost-efficient, hardware-ac

J-Zero: Unified Challenger--Solver--Judge Co-Evolution from Zero Data

👍 35

08/27 08:00

Self-evolving language models have recently emerged as a promising path toward superintelligence, with the advantage of reducing the cost of human supervision. While considerable progress has been made in verifiable domains, self-evolution in unverifiable domains remains substantially less explored.

Beyond Data Scaling: Representation-Centric Continued Pre-training for Vision-Language-Action Models

👍 60

08/27 08:00

Scaling robot data is crucial for building generalist Vision-Language-Action (VLA) models, yet robot trajectories are harder to scale than web-scale image-text data because embodied collection is costly and sparsely covers the physical world. This makes representation quality a central bottleneck: u

Code as Worlds: Agentic Discovery of Executable World Representations for Physical Reasoning

👍 43

08/27 08:00

Physical understanding and reasoning depend on forming compact and generalizable representations of the world. While modern vision-language models can recognize and explain diverse physical events, they often lack explicit representations of the underlying mechanisms-such as object states, physical

CritICL: Inference-Time Weak-to-Strong Generalization from Small Language Model Failure Modes

👍 9

08/27 08:00

Recent advances in inference-time scaling have significantly improved the reasoning performance of large language models (LLMs). However, these methods typically rely on repeated generation or external verification. To address this limitation, we introduce CritICL, a novel inference-time framework t

PILOT in the Loop: Live Self-Improvement for Long-Horizon Agents

👍 30

08/27 08:00

Long-horizon agent runs generate experience that can improve both the current run and future work. Most self-improvement methods process this experience only after execution ends, so they cannot redirect the active run or immediately apply and validate lessons learned from it. We argue that self-imp

GGSS: Geodesic-Gated Spherical Steering for Inference-Time Debiasing of Generative Vision-Language Models

👍 3

08/26 12:58

Generative vision-language models (VLMs) are increasingly used in human-centered settings, yet they can produce demographically biased outputs even when images differ only in controlled attributes such as perceived race or gender. However, existing inference-time debiasers were largely designed for

Paint What You See: Benchmarking Dexterous Visual Tool Use in Multimodal Agents

👍 14

08/26 08:00

Evaluation is shifting from static QA toward agentic settings where models act through external tools. We identify a critical yet underexplored capability within this space - dexterous visual tool use: fine-grained, closed-loop parameterized visual action in which models infer tool parameters from v

LMSM: LLM Security Framework Inspired by Linux Security Modules

👍 3

08/26 08:00

Large language models (LLMs) are increasingly deployed with layered defenses, yet malicious prompts can still bypass them. Interpretability methods can expose model-internal signals along the generation path that could inform enforcement, but these signals are not security controls by themselves. De

TacForcing: Streaming Action Generation with Execution-Time Tactile Feedback

👍 5

08/26 08:00

Contact-rich manipulation requires adapting to contact states that can evolve substantially within an action horizon. However, chunk-based vision-language-action models predict complete action chunks from observations collected before execution, leaving tactile conditioning stale during execution. E

Agentic Game Development as a Verifiable Trajectory Data Engine for Scaling World Models

👍 185

08/26 08:00

A common strategy for scaling world models is to train on more crawled video with more compute. We argue that this strategy is inefficient: scaling world models also requires a recursive data engine that offers grounded reward signals. The success of code agents illustrates why this matters. As code

Procedura: Agentic 3D Modeling with Procedural Control

👍 13

08/26 08:00

Native 3D generators now recover impressive mesh geometry from a single image. However, a dense mesh stays soft where a machined object should be sharp, it carries no part decomposition, and it exposes no parameter a user could edit. To address this, we explore the paradigm of 3D shape as code, leve

StarHarness: Evolving Harnesses with Stratified Search for Enterprise Environments

👍 9

08/25 08:00

We present StarHarness, a framework for evolving environment-specific agent harnesses while keeping model weights fixed. The evolved harness can include prompt and task framing, tool interfaces, skills, MCP-backed providers, subagent structure, and agent-loop configuration. StarHarness constructs a

StepGuard: Learning Step-Level Guardrails with Scalable Supervision and Safety-Utility Balancing

👍 14

08/25 08:00

LLM-based agents can interact with external environments through tool invocation, but this capability also introduces security risks such as file modification, information leakage, and unauthorized actions. Existing guardrails often evaluate completed trajectories, leaving pre-execution monitoring o

PonderPounce: A Pretrained MLLM as an Episode Context Engine for Robot Control

👍 11

08/25 08:00

Multimodal large language models (MLLMs) can integrate long visual histories, reason under partial observability, and infer behavior from a few examples. Yet vision-language-action (VLA) models generally inherit pretrained representations without using this contextual capacity as episode memory. Mem

Language Chain in Alignment: Cross-lingual Ranking Preference Optimization

👍 4

08/24 19:54

The alignment of Large Language Models heavily relies on English-centric high-quality preference data, which often leads to suboptimal performance in other languages. In this paper, we propose Cross-lingual Ranking Preference Optimization~(CRPO), a novel framework that leverages robust preference kn

Act with Intent: Distilling Behavior Intent for Vision-Language-Action Models

👍 24

08/24 08:00

Vision-Language-Action (VLA) models can turn multimodal context into robot actions, but their action decoders are still trained largely by behavior cloning. This supervises which motor command was demonstrated while leaving implicit the local objective served by the behavior under the instruction. F

Training, learning and inference: unified dynamics of neural systems

👍 3

08/21 08:00

We define an atomic generation fact f=(u,tau,omega,z;rho), recording the origin, realized transformation, concrete occurrence, generated result and relation role. Compiled into a Generation-Fact Graph (GFG), these facts provide an AI-native, compilable scientific fact substrate preserving generation

DART-SD: Diamond-topology Aware Retrieval and Tuning for Self-Distillation of Multi-Turn Tool-Calling Agents

👍 60

08/19 08:00

Equipping Large Language Models (LLMs) with multi-turn tool-calling capabilities is essential for building autonomous agents. However, progress is fundamentally limited by the reliance on full-length trajectory imitation. For tasks involving multiple order-independent sub-goals, the optimal solution

Lost in Compression: A Controlled Cross-Lingual Audit of Extractive Prompt Compressors

👍 2

07/27 08:00

Extractive prompt compression promises to cut LLM inference costs by removing low-information tokens, and learned compressors such as LLMLingua-2 report strong results on English benchmarks. Most other languages already pay a token premium: the same content costs 1.3-1.8x more tokens than in English

Department of War Launches OpenAI's ChatGPT Mil on GenAI.mil

@DoWCTO · 104.9K 粉丝 · 86.1K 阅 · 663 赞 · 139 转

08/31 22:25

The Department of War today launched OpenAI's ChatGPT Mil, marking it as the next frontier AI capability housed on GenAI.mil, the Department's premier generative AI platform. This

中文介绍 国防部在GenAI.mil平台推出OpenAI的ChatGPT Mil，标志着下一代AI能力的到来。

Agentic Engineering Setup (after 2,000+ hours)

@DavidOndrej1 · 67.4K 粉丝 · 68.7K 阅 · 513 赞 · 40 转

08/31 21:59

Over the past 3 years, I've spent well over 2,000 hours coding with AI, and I've personally interviewed some of the most productive people in the space of Agentic Engineering. Below is my full Agentic

中文介绍 分享2000小时AI编码经验，探讨Agentic Engineering领域的生产者访谈。

Grok Bot for Engineering

@lingxi · 7.0K 粉丝 · 64.4K 阅 · 629 赞 · 44 转

09/01 02:31

I’m a SpaceXAI engineer building Grok Bot with Grok Bot. Think of Grok Bot as a highly capable engineering intern, with its own computers, that can manage coding agents and learn from how you work. It

中文介绍 介绍Grok Bot，一款具备工程能力的AI实习生，可管理编码代理并学习用户工作方式。

Department of War Launches Starshield AI's Grok for Government on GenAI.mil

@DoWCTO · 104.9K 粉丝 · 57.5K 阅 · 924 赞 · 174 转

08/31 23:09

The Department of War today launched Starshield AI's Grok for Government, expanding the suite of frontier AI capabilities on GenAI.mil, the Department's premier generative AI platform.

中文介绍 国防部在GenAI.mil平台推出Starshield AI的Grok for Government，扩展前沿AI能力。

Terrible advice for software engineers

@steveruizok · 55.1K 粉丝 · 54.6K 阅 · 515 赞 · 35 转

08/30 01:12

I've recently started getting served outrage content videos from software engineers upset about AI coding. I want to share my thoughts about this content. I want you encourage you to watch the videos,

中文介绍 分析软件工程师对AI编码的愤怒内容视频，分享个人观点。

The Math of RL: From Policy Gradients to PPO and GRPO

@harshbhatt7585 · 3.1K 粉丝 · 48.0K 阅 · 500 赞 · 61 转

08/30 19:14

Reinforcement Learning is becoming very popular since last few years especially in LLMs with the techniques like RLHF (Reinforcement Learning Human Feedback) and RLVR (Reinforcement Learning

中文介绍 探讨强化学习在LLMs中的应用，从策略梯度到PPO和GRPO。

A Lagrangian View of Flow Matching

@docmilanfar · 115.8K 粉丝 · 26.0K 阅 · 512 赞 · 56 转

08/31 12:36

This is a pedagogical post meant to offer an intuitive, bottom-up perspective on Flow Matching. If you have ever wondered why some models require hundreds of iterative steps to generate an image while

中文介绍 从拉格朗日视角解读Flow Matching，提供直观的图像生成模型迭代步骤解释。

The Hugging Face hack could indicate cultural issues at OpenAI

MIT Tech Review AI · 09/01 02:00

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. By now you’ve probably heard about last month’s major AI security incident, in which OpenAI agents escaped their sandbox and hacked into the AI platform Hugging F

中文介绍 OpenAI近期AI安全事件可能暴露了其文化问题。

Polimill builds Japan's next-generation public AI infrastructure

OpenAI News · 08/31 15:00

Polimill uses OpenAI GPT models and Codex to help municipalities search and use administrative knowledge while accelerating development.

中文介绍 OpenAI的Polimill项目利用GPT模型助力日本市政知识搜索和开发加速。

A milestone in expanding access to AI

OpenAI News · 08/31 12:00

ChatGPT Ads reaches $1 billion in annualized revenue run rate and expands globally, supporting broader access to AI through free and affordable options.

中文介绍 ChatGPT广告年收入预计达10亿美元，全球扩展以促进AI普及。

OpenAI Cursor split 💔, self-improving AI 🧠, GPT-6 Astra 🚀

TLDR AI · 08/31 08:00

中文介绍 OpenAI Cursor分裂，自我改进AI和GPT-6 Astra发布。

[AINews] OpenAI shuts off Cursor

Latent Space · 08/29 13:11

Elon v Altman has a real consequence.

中文介绍 OpenAI关闭Cursor服务，Elon Musk与Altman的分歧产生影响。

[AINews] OpenAI to reach AGI bar by end-2026

Latent Space · 08/28 15:12

It’s Time. We’re in the Endgame now.

中文介绍 OpenAI预计到2026年底达到通用人工智能（AGI）里程碑。

Our decision on Cursor following its acquisition by SpaceX

OpenAI News · 08/28 14:00

Our decision to wind down our contract providing OpenAI models to Cursor following its acquisition by SpaceX.

中文介绍 OpenAI决定在SpaceX收购Cursor后停止提供模型服务。

Supporting Thailand’s next generation of AI startups

OpenAI News · 08/28 10:00

OpenAI and Thailand’s MHESI launch an eight-week accelerator helping 10 health, wellness, and education startups turn AI prototypes into trusted products.

中文介绍 OpenAI与泰国MHESI合作，支持AI初创企业发展。

The Open ASR Leaderboard Adds Its First Global South Language

Hugging Face Blog · 08/28 08:00

中文介绍 Open ASR排行榜新增首个全球南方语言。

Gemini Omni 1.1 🎬, Cohere Parse 📄, Codex persistent mode 👨‍💻

TLDR AI · 08/28 08:00

中文介绍 Gemini Omni 1.1、Cohere Parse和Codex持久模式发布。

Gemini Omni 1.1 Flash lets you build with more control

DeepMind Blog · 08/28 00:11

中文介绍 DeepMind发布Gemini Omni 1.1 Flash，增强构建控制。

Piloting the world's first double-blind AI evaluations

DeepMind Blog · 08/27 20:59

Piloting the world's first double-blind AI evaluations

中文介绍 DeepMind进行全球首次双盲AI评估试验。

Better answers, broader thinking: What students gain from ChatGPT and critical-thinking training

OpenAI News · 08/27 17:00

A randomized study of more than 1,000 students examines ChatGPT, critical thinking, originality, and student performance on a real-world university assignment.

中文介绍 研究显示，ChatGPT和批判性思维训练对学生表现有积极影响。

Expanding OpenAI’s presence in Brazil

OpenAI News · 08/27 11:00

OpenAI is expanding its presence in Brazil, deepening engagement with developers, businesses, and communities to support AI adoption across the country.

中文介绍 OpenAI在巴西扩大影响力，深化与开发者、企业和社区的互动。

[AINews] NVIDIA buys HuggingFace for $13B, as OpenAI publishes their HF incident retro

Latent Space · 08/27 09:50

Open Source wins!

中文介绍 英伟达以130亿美元收购HuggingFace，开源社区获得胜利。